Xavier Holt deleted In_terms_of_results_our__.tex  over 8 years ago

Commit id: e025fa2a6bd4a23f49e62926c04da92b5feb2eb0

deletions | additions      

         

In terms of results, our Bayesian logit model performed consistently less accurately and more slowly. In terms of sanity check, our MLE estimated is correctly situated around the peak of each of our posterior distributions. This is unsurprising, as we have ensured that our $\lambda$ parameters align in terms of posterior distribution/goal function. While there is a clear Markovian behaviour present in the sampling process, the formulation of our algorithm guarantees us convergence in distribution, at least in the limit. As such we do not attribute the model's poor behaviour to this.         

Each_cell_is_the_average__.tex  figures/WEIGHTS DIST/WEIGHTS DIST.png  figures/SamplingCONV/SamplingCONV.png  In_terms_of_results_our__.tex  Gridsearch__.md  Gridsearch_is_included_here_only__.tex  figures/FN/FN.png